little more than that of the Springhead example. Though much oxidized, it yet retains the undulated border, and it is perforated by seven nail-holes of very large size, with the oblong socket to lodge the nail-head. Three of the holes are on each side, and one in the centre of the toe has doubtless been intended to act like the modern toe-clip, and prevent the shoe from being driven back. This feature in these antique shoes is very rare, indeed this is the only instance in which I have been able to trace it. The aperture for the shank of the nail, instead of being nearly circular, as in the Springhead shoe, is quadrilateral, and of immense size, in proportion to the shoe (3/8 ths long, by 3/16 ths wide). One of the nails yet remains in the shoe, but the head is much worn; though sufficient is left to prove that it was of the flattened, high, and wide T pattern. The shank is almost square like a carpenter's nail, and fills the hole; and at a distance of only 1/4 inch from the foot surface of the shoe it bends suddenly forward as if to form a clench on the outside of the hoof. The excessive thickness of the nail, and the very short hold it had of the hoof, are easily accounted for. The shoe has evidently been for the near (left) fore foot, and the inner branch towards the heel is narrower than the outer one; it shows faint traces of a calkin, but the outer heel has a well-defined calkin formed by doubling over the extremity, as in the other specimens of this period, though this has been more clumsily done than in some of those we have noticed. The foot-surface is slightly concave from the outer to the inner rim.
In the large collection of undoubted Roman remains brought to light in this castra, are three spurs of antique
